Kreuzberg Document Extraction

Kreuzberg extracts text, tables, metadata, and images from 91+ file formats. Use it for document processing, OCR, batch extraction, structured LLM extraction, and embeddings.

Run kreuzberg --help for all commands, kreuzberg <command> --help for full flag reference.

Core Usage

# Single file → stdout (text) or JSON
kreuzberg extract document.pdf
kreuzberg extract document.pdf --content-format markdown --format json

# Batch
kreuzberg batch *.pdf --content-format markdown

# Detect MIME type
kreuzberg detect unknown-file